Mild Protease Treatment as a Small-Scale Biochemical Method for Mitochondria Purification and Proteomic Mapping of Cytoplasm-Exposed Mitochondrial Proteins Francesca Forner,§ Edgar A. Arriaga,§,† and Matthias Mann*,§ Max Planck Institute for Biochemistry, Department of Proteomics and Signal Transduction, Am Klopferspitz 18, D-82152 Martinsried, Germany, and Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455 Received July 21, 2006
Because of its importance in basic biology and medicine, great efforts are being devoted to unraveling of the genuine mitochondrial proteome, which is the dynamic protein complement that the organelle uses to maintain its structure and functionality. Several proteomic investigations have now clearly shown that all the purification approaches we have at our disposal suffer from the problem of co-purification; therefore, it is very difficult to distinguish novel mitochondrial proteins from those that are just contaminants of the preparation. The question is further complicated by the fact that the mitochondrial proteome depends on the tissue source. Density gradient centrifugation is the most widespread purification method for obtaining highly pure mitochondrial fractions. The main disadvantage of these methods is the low yield of purified mitochondria that precludes their use in low-scale purifications. Here, we have treated small aliquots of crude mitochondria from mouse liver and from cultured hepatocytes (HEPA1-6) with trypsin under mild proteolysis conditions and have evaluated the suitability of this reaction as a small-scale purification approach. The protease removed several cytoplasmic and endoplasmic reticulum proteins, together with a fraction of mitochondrial proteins that we hypothesize to be associated with the cytosolic face of the outer mitochondrial membrane. The peculiar topology of these mitochondrial proteins could be indicative of their functional roles. Finally, our study represents an application of advanced mass spectrometry technology to the evaluation of biochemical approaches for the treatment of mitochondria. Keywords: mitochondria • outer membrane • small-scale • purification • proteolysis • mass spectrometry • orbitrap • organellar proteomics
Introduction Mitochondria are known to have key roles in several metabolic pathways,1-4 in cancer, and in aging.5-7 Moreover, mitochondrial mutations are responsible for several encephalomyopathies,8 and mitochondrial abnormalities have been found in Parkinson,9 Alzheimer,10 and Huntington11 diseases. For these reasons, mitochondria represent an important target for proteomic studies.12-16 Biochemical methods for the characterization of these organelles are being enormously enhanced by the rapidly developing capabilities of protein mass spectrometry (MS) and are an area of interest in our laboratory.17,18 The capacity of localizing novel proteins to specific organelles in the cell with high confidence has major functional implications that go far beyond simple protein mapping. For example, novel members of the mitochondria proteome may immediately * To whom correspondence should be addressed. Max Planck Institute for Biochemistry, Department of Proteomics and Signal Transduction, Am Klopferspitz 18, D-82152 Martinsried, Germany. Tel: +49 89 8578 2557. Fax: +49 89 8578 2219. E-mail:
[email protected]. § Max Planck Institute for Biochemistry. † University of Minnesota. 10.1021/pr060361z CCC: $33.50
2006 American Chemical Society
suggest novel roles of this multifunctional organelle which can be tested experimentally. However, the results of studies that characterize subcellular organelles are strongly influenced by the purity of preparations. Experimentally, cells and tissues are usually disrupted and organelles enriched by means of density centrifugation gradients. Extracted proteins are then characterized through antibody detection and/or mass spectrometry. Recently, a new approach named protein correlation profiling19 was applied to provide a subcellular localization for 1400 proteins in the mouse liver.20 Despite the established biochemical methods for the isolation of mitochondria from several animal tissues and the long time that density centrifugation gradients have been used for the removal of contaminant proteins,21 highly sensitive protein detection with MS has lately demonstrated that it is very difficult to determine novel mitochondrial proteins with high confidence.17,18,20 Moreover, the use of density gradients (such as sucrose or Percoll) for small-scale separations of mitochondria yields insufficient amounts of purified material. This problem is exacerbated if mitochondria are to be obtained from cultured cells where usually far less mitochondria can be Journal of Proteome Research 2006, 5, 3277-3287
3277
Published on Web 11/10/2006
research articles
Figure 1. Experimental approach for evaluating the protease treatment of crude mitochondria isolated from tissue and from cultured cells. Crude mitochondria were isolated from the mouse liver and from cultured hepatocytes (HEPA 1-6). Two purification approaches were performed in parallel. One aliquot was treated with trypsin in controlled conditions, whereas a second aliquot was purified through a Percoll density-gradient centrifugation. Mitochondria obtained after the two treatments were solubilized and subjected to in-solution digestion. Proteins identified in the two samples were compared based on sequence alignment (Blast).
isolated. Yields are also very dependent on the cell type, and vary considerably from tissue to tissue. For example, muscle cells and skeletal muscle tissue are in principle rich sources of mitochondria; however, isolation of substantial amounts of purified mitochondria is far from trivial.18 Here, we introduce an alternative treatment of crude mitochondrial fractions that resulted in the removal of the main protein contaminants without damaging the mitochondrial structure. The procedure has been evaluated for small-scale preparations. We have treated crude mitochondrial preparations with trypsin in controlled reaction conditions. Trypsin has a high cleavage specificity22 and is quite tolerant to reaction buffer composition and to low temperatures.23 We have first used a normal-scale approach based on MS to evaluate the effectiveness of the protease treatment as an alternative mild purification approach and compared it to the well-established density centrifugation gradients (Figure 1). The procedure provided comparable results in terms of purity but much superior results in terms of yield when applied to small-scale mitochondrial fractions isolated from hepatocytes in cell culture. We suggest that the procedure can be applied for the purification of low-scale preparations or, alternatively, as a precleaning method before a density gradient. Protease treatment has long been used to study the import of protein into mitochondria indeed, proteins that show low sensitivity to proteases are considered to be protected and, therefore, imported. In this respect, the proteins that were affected by our protease treatment provided interesting information on the composition of the outer mitochondrial membrane facing the cytosolic bulk. The outer mitochondrial membrane has received increasing attention, but while it has been investigated in depth with respect to the import machinery complexes, few other functions have been described in detail. The outer membrane allocates the translocase complex 3278
Journal of Proteome Research • Vol. 5, No. 12, 2006
Forner et al.
(TOM) for the import of mitochondrial proteins24 and the VDAC complex25 which is involved in the import of ions and metabolites, as well as several integral membrane proteins with R-helical and β-barrel transmembrane segments. On the other hand, the identity and the functional relevance of the pool of peripheral membrane proteins interacting with the outer membrane remain elusive. In recent work, a mitochondrial surface fraction was described based on the peptides identified in the postcentrifugation supernatant of purified yeast mitochondria treated with trypsin.26,27 Several proteins of the internal compartments were identified, and the authors hypothesized an accumulation of pre-proteins at the outer membrane. Although here we have focused on the protease treatment mainly as a purification method, the approach also enabled the identification of a class of mitochondrial proteins that were effected by trypsin digestion. The apparent localization of these proteins on the cytoplasmic side of the outer membrane is discussed in relation to the published literature. Besides its direct involvement in organellar fission/fusion in response to changes in cellular conditions, the cytosolic face of the outer membrane can be thought of as the organelle’s window on the outside cellular environment. Therefore, it is reasonable to expect that it is also involved in other signaling pathways targeting mitochondria. We hypothesize that the proteins that were removed by trypsin may have a functional role related to their particular location.
Experimental Section Materials. Trypsin was from Promega. Digitonin was from Sigma-Aldrich. Anti-calreticulin is a mouse monoclonal antibody from BD Transduction Laboratories (C41720). Anti-porin is a mouse monoclonal antibody from Invitrogen (A31855). Anti-cytochrome oxidase subunit I is a mouse monoclonal antibody from Invitrogen (A6403). Isolation of Mitochondria from Liver Tissue and from Cells. Three months old C57BL/6 female mice were starved overnight with water ad libitum. After sacrificing the animal by cervical dislocation, the liver was quickly removed and washed three times in 250 mM sucrose, 10 mM Tris-HCl, and 0.1 mM EGTA (pH 7.4) supplemented with protease inhibitors (Roche). Crude mitochondria were isolated as described18 and either purified in a Percoll density centrifugation as described18 or treated with trypsin (see below). HEPA1-6 cells were cultured in 15 cm dishes in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum at 37 °C under an atmosphere of 5% CO2 and 95% air. The cells were harvested with trypsin for 5 min at 37 °C. All subsequent operations were performed on ice or at 4 °C. Cells were washed twice and resuspended in 10 mL of 180 mM sucrose, 10 mM Tris-HCl, and 0.1 mM EGTA (pH 7.4) supplemented with protease inhibitors cocktail (Roche). After 10 strokes of the tight-fitting Dounce homogenizer, 10 µL of 1% digitonin in DMSO was added. Twenty additional strokes were applied, and the cell disruption was monitored under the microscope with Trypan blue. The cell homogenate was aliquoted in 2 mL tubes and centrifuged for 5 min at 600g. Crude mitochondria were pelleted by centrifuging the supernatant for 15 min at 10 000g. Mitochondria were resuspended in 250 mM sucrose, 10 mM Tris-HCl, and 0.1 mM EGTA (pH 7.4) and centrifuged for 10 min at 7000g. Mitochondria were then purified as described for the liver-isolated mitochondria. Citrate synthase-specific activity was determined as described.28
research articles
Small-Scale Purification of Mitochondria
Protease Treatment of Mitochondria. Aliquots corresponding to 100 µg of crude mitochondrial proteins were resuspended to a final volume of 100 µL with 250 mM sucrose, 10 mM TrisHCl, and 0.1 mM EGTA (pH 7.4). After addition of DTT to a final concentration of 2 mM, the suspension was incubated on ice for 15 min. Only where indicated, iodoacetamide was added at a final concentration of 10 mM and incubated on ice for 15 min in the dark. Trypsin was added to an estimated protein/enzyme ratio of approximately 50:1. The reaction mixture (pH ) 7) was incubated overnight at 4 °C. Proteasetreated mitochondria were centrifuged for 10 min at 7000g, resuspended in the same buffer supplemented with the protease inhibitor cocktail, and centrifuged two more times. Aliquots corresponding to 10 µg of total protein (Bradford assay) were diluted to 25 µL with dilution buffer (6 M urea, 2 M thiourea, and 10 mM Hepes, pH 8), reduced, alkylated, and digested to peptides with trypsin and endopeptidase LysC as described.18 NanoLC-MS/MS, MASCOT Database Search, and Data Analysis. Resulting tryptic peptides were desalted and concentrated on reversed-phase C18 StageTips.29 Peptides were resuspended in 10 µL of 3% acetonitrile, 1% trifluoroacetic acid, and 0.5% acetic acid before injection. Liquid chromatography was performed on a 15 cm fused silica emitter (75 µm i.d. from Proxeon Biosystems, Odense, Denmark) packed in-house with reversed-phase ReproSil-Pur C18-AQ 3 µm resin (Dr. Maisch GmbH, Ammerbuch-Entringen, Germany). The injection volume was 5 µL, and the flow rate was 250 nL/min after a tee splitter, as indicated by flow measurement device. The experiments were performed on an Agilent 1100 nanoflow system connected to an LTQ Orbitrap mass spectrometer (Thermo Electron, Bremen, Germany) equipped with a nanoelectrospray ion source (Proxeon Biosystems, Odense, Denmark). The peptide mixtures were injected onto the column with a flow of 500 nL/min and subsequently gradient-eluted with a flow of 250 nL/min from 5% to 40% CH3CN in 0.5% acetic acid. Gradients were 140 min long. The mass spectrometer was operated in the data-dependent mode to automatically switch between orbitrap-MS and ion trap-MS/MS (MS2) acquisition. Survey full scan MS spectra (from m/z 350 to 1550) were acquired in the orbitrap with resolution R ) 60 000 at m/z 400 and accumulation to a target value of 1 million charges. The five most intense ions in each full scan were sequentially isolated for fragmentation in the linear ion trap using collisionally induced dissociation at a target value of 10 000 charges. The resulting fragment ions were recorded in the ion trap with resolution R ) 15 000 at m/z 400. For accurate mass measurements, the lock-mass option was enabled in both MS and MS/ MS modes as described.30 Target ions already selected for MS/ MS were dynamically excluded for 30 s. General mass spectrometric conditions were electrospray voltage, 2.3 kV; no sheath and auxiliary gas flow; ion transfer tube temperature, 150 °C; collision gas pressure, 1.3 mTorr; normalized collision energy, 30% for MS2. Ion selection threshold was 250 counts for MS2. An activation q ) 0.25 and activation time of 30 ms was applied for MS2 acquisitions. Proteins were identified by automated database searching (Mascot Daemon, Matrix Science) against an in-house curated version of the IPI mouse (Mus musculus) protein sequence database. This database was complemented with frequently observed contaminants (porcine trypsin, Achromobacter lyticus lysyl endopeptidase, and human keratins). Search parameters specified an initial MS tolerance of 10 ppm and an MS/MS
tolerance at 0.4 Da and Trypsin/P + DP specificity allowing for up to 2 missed cleavages. Carbamidomethylation of cysteine was set as a fixed modification, and oxidation of methionines and N-protein acetylation were allowed as variable modifications. Because of the high mass accuracy, the 99% significance threshold in the IPI mouse database search was a Mascot score of 27. Acceptance criteria for protein identifications were at least two sequenced nonredundant peptides (“unique peptides”) with mass accuracy better than 5 ppm and at least one of them with Mascot score higher than 27. Therefore, peptides differing only by their charge were considered nonunique and were treated as a single identification. In total, 4926 unique peptide sequences with more than 5 amino acids in length and mass accuracies better than 5 ppm were identified. The average absolute mass accuracy of the precursor peptides was 1.015 ppm. For BLASTing (ftp://ftp.ncbi.nih.gov/blast/executables/, BLOSUM62 scoring matrix) protein sequences from two different experiments, we considered sequence alignments of at least 80% to be indicative of the protein presence in both. This approach was based on the observation that the simple comparison of the protein identification codes (IDs) was misleading due to the presence of redundant annotations and isoforms in the database. The reason for having chosen an 80% alignment instead of 100% was that, although in some cases the same peptides were sequenced in the two samples, the database search matched them to slightly different protein sequences (splice isoforms, fragment sequences, etc.). The 80% alignment was a reasonable compromise to have the correct protein sequences matched in these situations. In these cases, the reported protein sequence was that of the reference file chosen as database for the Blast analysis.
Results and Discussion Mouse Liver Mitochondria Treated with Trypsin. Crude and highly purified fractions of liver mitochondria were incubated with trypsin at 1:50 enzyme/substrate ratio and at a temperature of 4 °C. Whereas trypsin is not inhibited at 4 °C, the activity of many other endogenous proteases which are active at physiological conditions should be decreased. Furthermore, it has been shown that phospholipids in the mitochondrial membranes maintain their bilayer configuration at 4 °C for several hours, but after few minutes at 37 °C, the in vivo lipid organization is lost.31 Therefore, we chose to perform the proteolysis at low temperature. Mitochondria resulting from the trypsin reaction should be devoid of the cytosol-exposed and proteins loosely associated to the outer mitochondrial membrane (Figure 2). To verify the intactness of the mitochondrial membranes after protease treatment, we measured the activity of the matrix enzyme citrate synthase over several hours in both the crude and the protease-treated mitochondrial fractions (Supporting Information Figure 1). Citrate synthase is the best-known marker enzyme to determine the damage done to mitochondria during preparation,28 and no activity should be detected in a preparation of completely intact mitochondria. The specific activity in both crude and proteasetreated mitochondrial aliquots was close to zero over several hours. Conversely, an aliquot of purposely denatured mitochondria showed significant activity. This clearly demonstrates the intactness of mitochondria after the overnight trypsin reaction. Antibodies directed against calreticulin (endoplasmic reticulum marker) and porin (outer mitochondrial membrane marker) showed that the treatment was effective in removing Journal of Proteome Research • Vol. 5, No. 12, 2006 3279
research articles
Forner et al.
Figure 2. Topology of the proteins accessible to mild protease treatment. Among the proteins that were removed by the protease treatment of crude mitochondrial preparations, we could distinguish the following classes: proteins that co-purify with mitochondria during the differential centrifugation in isotonic buffer (e.g., cytoplasmic proteins); proteins associated with other organellar membranes that partially sediment at 7000g (e.g., endoplasmic reticulum, lysosomes, and peroxisomes); mitochondrial proteins associated with the cytoplasmic side of the outer membrane; and mitochondrial membrane proteins with cytoplasm-exposed domains.
the first marker, but also affected the abundance of the outer mitochondrial membrane protein. In contrast, the level of the inner mitochondrial membrane protein cytochrome oxidase subunit I was not diminished by the protease treatment (Figure 3). To obtain a more thorough evaluation of the performance of the proteolysis reaction, we performed a large-scale comparison of the proteins in the trypsin-treated mitochondria with those in the crude mitochondrial preparation by means of highaccuracy/high-sensitivity MS. We determined which proteins were uniquely identified in the crude mitochondrial sample (and therefore removed by the proteolysis) by blasting the protein sequences against a database composed of the sequences identified in the trypsin-treated sample. Of these proteins, those identified with at least 2 unique peptides (see Experimental Section for a definition of unique) were classified according to their Swiss-Prot subcellular localization annotation (www.expasy.ch/) (Table 1). Peptide identification details for the proteins reported in Table 1 can be found in the Supporting Information Table 3. Several known contaminant proteins of mitochondrial preparations were identified in the crude mitochondrial sample, which is expected, as upon differential centrifugation, mitochondria partially co-sediment with other cell organelles. The list in Table 1 is not exhaustive due to the complexity of the analyzed sample and the finite dynamic range of mass spectrometry. Furthermore, the reported proteins represent those that were most sensitive to protease treatment. Nevertheless, the proteins listed in Table 1 represent a means of evaluation of the protease treatment with respect to the standard protocol. The highest percentage of removed proteins belongs to the cytoplasmic compartment, followed by mitochondrial, endoplasmic reticulum, and nuclear proteins. The mitochondrial distribution in eukaryotic cells usually shows a dense organelle accumulation in the perinuclear region and seems to imply physical attachment (Supporting Information Figure 2). Therefore, it is common to identify nuclear proteins in MS-analyzed mitochondrial preparations. In our hands, histones H2A and H2B appeared in the crude fractions but were removed by this treatment (H2B was identified with only one unique peptide in the crude fraction and therefore does not appear in Table 1). This indicates that these proteins most likely co-sediment with mitochondria during preparation of the crude 3280
Journal of Proteome Research • Vol. 5, No. 12, 2006
Figure 3. Effect of protease treatment on calreticulin, porin, and cytochrome oxidase. (Upper panel) The protease treatment decreased the amount of the endoplasmic reticulum protein calreticulin levels in the crude mitochondrial fraction and after the density-gradient centrifugation. (Middle panel) The treatment similarly decreased the cytoplasm-exposed part of the outer mitochondrial membrane protein porin. (Lower panel) The protease treatment did not affect levels of the inner mitochondrial membrane protein cytochrome oxidase. CM, crude mitochondria; CM + Try, crude mitochondria treated with Trypsin; GM, densitygradient-purified mitochondria; GM + Try, density-gradientpurified mitochondria treated with Trypsin.
fraction. This finding would also explain why histones and histone-like proteins are sometimes detected in mitochondrial preparations.32 As expectedsbecause of their association with the outer mitochondrial membranesseveral ribosomal proteins
research articles
Small-Scale Purification of Mitochondria
Table 1. Proteins That Were Removed by the Density-Gradient Centrifugation (G) and by the Trypsin Protease Treatment (T)a organelleb
Cytoplasm
IDc
protein name
1110002E23Rik protein 2700085E05Rik protein 1110039B18Rik protein ApoA-I binding protein Cis-retinol androgen dehydrogenase 1 Complement C3 precursor Flad1 Hypothetical EF-hand/S-100/ICaBP type calcium binding protein containing protein Hypothetical peptidase family M20/M25/M40 containing Hypothetical PPR repeats containing protein Hypothetical protein Hypothetical protein Hypothetical protein Gulo NADPH-dependent retinol dehydrogenase/reductase homologue Ociad1 protein Protein FAM82B Putative steroid dehydrogenase KIK-I Quinoprotein alcohol dehydrogenase structure containing protein Retrovirus-related Pol polyprotein LINE1 RIKEN cDNA 0710008K08 Serine carboxypeptidase 1 Similar to NADP-dependent malic enzyme Sitpec-pending protein Tarsl1 protein UDP-glycosyltransferase 3 family, A2 1 UPF0240 protein C6orf66 homologue 14-3-3 protein gamma 260S ribosomal protein L3, similar to 40S ribosomal protein S13 40S ribosomal protein S26, similar to 40S ribosomal protein S3 40S ribosomal protein S3a 40S ribosomal protein S5a 40S ribosomal protein S6 40S ribosomal protein S8 40S ribosomal protein S9 40S ribosomal protein SA 60S acidic ribosomal protein P2 60S ribosomal protein L10 60S ribosomal protein L10a 60S ribosomal protein L13a 60S ribosomal protein L18 60S ribosomal protein L18a 60S ribosomal protein L23a 60S ribosomal protein L26 60S ribosomal protein L27 60S ribosomal protein L28 60S ribosomal protein L3, similar to 60S ribosomal protein L5 60S ribosomal protein L6 60S ribosomal protein L7 60S ribosomal protein L7a 1, similar to 60S ribosomal protein L7a, similar to 60S ribosomal protein L9 Actin, alpha skeletal muscle Actin, cytoplasmic 1 Alcohol dehydrogenase Arginase-1 Arginyl-tRNA synthetase-like Betaineshomocysteine S-methyltransferase Bifunctional coenzyme A synthase C-1-tetrahydrofolate synthase Catechol O-methyltransferase Ciliary dynein heavy chain 5 Elongation factor 1-alpha 1 Elongation factor 2 Fatty acid-binding protein, liver Fructose-bisphosphate aldolase B Glutathione S-transferase Yc Glycerol-3-phosphate dehydrogenase Guanine nucleotide-binding protein beta subunit 2-like 1 Heat shock protein 1 Heat shock-related 70 kDa protein 2
G
Q9CQN3 Q7TMP4 Q8R237 Q8K4Z3 O54909 P01027 Q3TSN6 Q9D3P1
x x
Q8C165 Q3UKH5 Q9DCZ4 Q8BNY5 Q8K156 Q9D2U3 Q922M2 Q9DCV4 O70503 Q8C7X2 XP_920398 Q9D7G7 Q99J29 XP_923436 Q91V53 Q922A3 Q8JZZ0 Q9D1H6 P61982 XP_620346 P62301 XP_898748 P62908 P97351 P62245 XP_125109 P62242 Q6ZWN5 P14206 P99027 Q6ZWV3 P53026 P19253 P35980 P62717 P62751 P61255 P61358 P41105 XP_620346 P47962 P47911 P14148 XP_926227 XP_892491 P51410 P68134 P60710 Q9JII6 Q61176 Q3U186 O35490 Q9DBL7 Q922D8 O88587 Q8VHE6 P10126 P58252 P12710 Q91Y97 P30115 P13707 P68040 Q3UIQ7 P17156
x x x x x x
x x x x
T
x x x x
x
x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x
x x x x x x x x x x x x x x x
x
x
x x x x x x X x x
peptidesd
2 2 2 3 2 2 2 2 2 3 2 2 7 2 2 5 3 3 2 3 4 10 2 2 5 2 2 3 3 2 6 4 4 3 4 5 4 4 2 4 5 3 3 2 4 2 3 3 4 4 4 4 3 4 5 5 2 3 2 5 3 5 6 2 4 7 2 2 2 6 4 2 2
Journal of Proteome Research • Vol. 5, No. 12, 2006 3281
research articles
Forner et al.
Table 1 (Continued) organelleb
Cytoplasm/Mitochondria Cytoplasm/Nucleus Peroxisomes
Nucleus Early Endosomes Endoplasmic Reticulum
Endoplasmic Reticulum Endoplasmic reticulum/nucleus ER and Golgi ER and/or Golgi Membrane
3282
protein name
IDc
G
T
peptidesd
Hexaprenyldihydroxybenzoate methyltransferase Hypothetical GTP-binding elongation factor containing protein Hypothetical ribosomal protein L21 Isocitrate dehydrogenase 1 Maleylacetoacetate isomerase NAD-dependent deacetylase sirtuin-3 NADP-dependent malic enzyme Peroxiredoxin 1 Peroxiredoxin-4 Quinone oxidoreductase Ras-related protein Rab-10 Ras-related protein Rab-1B Ribosomal protein L38, similar to Rps16 protein Smooth muscle gamma-actin Superoxide dismutase T-complex protein 1, epsilon subunit T-complex protein 1, zeta subunit Stomatin-like protein 2 Staphylococcal nuclease domain-containing protein1 Vigilin Dehydrogenase/reductase SDR family member 4 Peroxisomal 2,4-dienoyl-CoA reductase Peroxisomal carnitine O-octanoyltransferase Peroxisomal membrane protein 2 Peroxisomal membrane protein 4 Peroxisomal sarcosine oxidase Histone H2A.291.A Proliferation-associated protein 2G4 Ras-related protein Rab-5B 3-beta-hydroxysteroid-delta(8),delta(7)-isomerase Acyl-CoA desaturase 2 Arylacetamide deacetylase Calnexin precursor Cytochrome b5 Cytochrome P450 2A5 Cytochrome P450 3A11 Cytochrome P450 3A13 Cytochrome P450 4A10 Cytochrome P450 4V3 Cytochrome P450 7B1 Cytochrome P450 8B1 Cytochrome P450, 2d22 Cytochrome P450, family 2a12 Cytochrome P450, family 2d22 Ectonucleoside triphosphate diphosphohydrolase 5 Glucose-6-phosphatase Hypoxia up-regulated 1 Mannosyl-oligosaccharide glucosidase Microsomal triglyceride transfer protein large subunit Protein transport protein Sec61, alpha 1 Ribosome-binding protein 1 Thioredoxin domain-containing protein 5 Translocon-associated protein alpha Translocon-associated protein delta UDP-glucose:glycoprotein glucosyltransferase 1 Protein LYRIC Neutral alpha-glucosidase AB Fatty-acid amide hydrolase Camello-like 2 Ezrin Flavin containing monooxygenase 5 Growth and transformation-dependent protein Hypothetical protein Hypothetical protein Nicalin Protein transport protein SEC61, gamma Ras-related protein Rab-1B Similar to integral membrane homologue Solute carrier family 21 member 1 Suppressor of actin mutations (Sac1p) Vesicle-associated membrane protein-associated protein A Vesicle-associated membrane protein-associated protein B
Q8BMS4 Q8C3X4 Q9D1N9 O88844 Q9WVL0 Q8R104 P06801 P35700 O08807 P47199 P61027 Q9D1G1 XP_895159 Q641N3 Q61852 P08228 P80316 P80317 Q99JB2 Q78PY7 Q8VDJ3 Q99LB2 Q9WV68 Q9DC50 P42925 Q9JJW0 Q9D826 P10812 P50580 P61021 P70245 P13011 Q99PG0 P35564 P56395 P20852 Q64459 Q64464 O88833 Q9DBW0 Q60991 O88962 Q91W87 Q8VCW9 Q91W87 Q9WUZ9 P35576 Q9JKR6 Q80UM7 O08601 P61620 Q99PL5-1 Q91W90 Q9CY50 Q62186 Q6P5E4 Q80WJ7 Q8BHN3-2 O08914 Q8C7E5 P26040 Q3UV28 Q9D6U8 Q8BMB6 Q4KKX5 Q8VCM8 P60060 Q9D1G1 Q3TDV1 Q9QXZ6 Q9EP69 Q9WV55 Q9QY76
x x x x x x x x x x x
x x x x
2 2 2 3 5 2 10 2 2 2 3 2 2 3 3 2 2 2 3 8 6 2 3 7 5 2 5 2 4 2 2 2 7 4 2 4 8 5 4 3 9 4 3 4 3 2 2 8 3 9 3 3 3 3 5 2 2 10 3 6 2 7 2 2 2 2 2 2 2 3 2 2 2
Journal of Proteome Research • Vol. 5, No. 12, 2006
x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x
x x x x x
x x x x x x x x x x x x x x x x x x
x
x
x
x x x
x x x x
research articles
Small-Scale Purification of Mitochondria Table 1 (Continued) organelleb
Golgi
Golgi/cytoplasm Mitochondria
Mitochondria
protein name
IDc
Aflatoxin B1 aldehyde reductase 2 Ras-related protein Rab-1A Membrane protein p24A Prostaglandin E synthase 2 28S ribosomal protein S16 28S ribosomal protein S34 28S ribosomal protein S30 39S ribosomal protein L46 39S ribosomal protein L12 2-amino-3-ketobutyrate coenzyme A ligase 4-aminobutyrate aminotransferase 1 2-oxodicarboxylate carrier Acad10 protein Acyl coenzyme A thioester hydrolase Acyl-CoA dehydrogenase family, 9 Adenine nucleotide translocator 1 Adenylate kinase isoenzyme 4 Adrenodoxin oxidoreductase AFG3-like protein 2 Argininosuccinate synthase ATP synthase f chain ATP synthase protein 8 ATP-dependent Clp protease ATP-binding subunit ClpX-like Carbonic anhydrase 5A Cysteine desulfurase Cytochrome c oxidase polypeptide VIIa-liver/heart Dehydrogenase/reductase SDR family, 1 Dynamin-like 120 kDa protein (Opa1) Elongation factor G1 Enoyl Coenzyme A hydratase domain containing 3 Fission 1 protein (Fis1) Frataxin Glutaredoxin-related protein 5 Glycerol-3-phosphate acyltransferase Grave disease carrier protein homologue Hypothetical aminoacyl-transfer RNA synthetases class-II/ Dihydrodipicolinate synthetase containing protein Lactamase, beta 2 Metaxin-2 Methylmalonyl-CoA epimerase Mimitin Mitochondrial glutamate carrier 1 MOCO sulfurase C-terminal domain 2 NADH dehydrogenase, 6 NADH-ubiquinone oxidoreductase, 1 NADH-ubiquinone oxidoreductase, 13 A NADH-ubiquinone oxidoreductase, 20 NADH-ubiquinone oxidoreductase, B14 NADH-ubiquinone oxidoreductase, B14.7 NADH-ubiquinone oxidoreductase, B17.2 NADH-ubiquinone oxidoreductase, B18 NADH-ubiquinone oxidoreductase, ASHI NADH-ubiquinone oxidoreductase, ESSS NADH-ubiquinone oxidoreductase SGDH Nucleoside diphosphate-linked moiety X 8 Ociad2 Peptidyl-tRNA hydrolase 2 Peripheral-type benzodiazepine receptor Peroxisomal lon protease homologue Protein C14orf159 homologue Pyruvate dehydrogenase E1, alpha Rhot1 Ribosomal protein L4 SA rat hypertension-associated homologue Sideroflexin-2 Translocase of outer membrane TOM22 Translocase of outer membrane TOM70 Tyrosyl-tRNA synthetase, probable Ubiquinol-cytochrome c reductase complex 7.2 kDa protein Zinc-binding alcohol dehydrogenase domain-containing protein 2
Q8CG76 P62821 Q9R0Q3 Q8BWM0 Q9CPX7 Q9JIK9 Q9D0G0 Q9EQI8 Q9DB15 O88986 P61922-1 Q8BZ09 Q9CRH8 Q9QYR9 Q8JZN5 Q8CFJ7 Q9WUR9 Q61578 Q8JZQ2 P16460 P56135 P03930 Q9JHS4 P23589 Q9Z1J3 P48771 Q99L04 P58281 Q8K0D5 Q99LV0 Q9CQ92 O35943 Q80Y14 Q8VCT2 Q8C0K5-1 Q9CY60 Q99KR3 O88441 Q9D1I5 Q59J78 Q9D6M3 Q922Q1 Q3TRR5 P03888 Q3TRR5 Q9DC70 Q9CQZ5 XP_155879 Q7TMF3 Q9CR61 Q9D6J5 O09111 Q9CQH3 Q9CR24 Q9D8W7 Q8R2Y8 P50637 Q8BK80 Q8BH86-1 P35486 Q8BLW3 Q8JZU9 Q91WI1 Q925N2 Q9CPQ3 Q9CZW5 Q8BYL4 Q8R1I1 Q8BGC4
G
x x x x x x x x x x x x x x
T
peptidesd
x x
3 2 2 5 3 2 3 2 3 5 14 2 6 2 5 2 2 2 2 10 2 2 4 3 2 3 3 2 3 2 2 2 2 5 2 3
x x
x x
x x x x x x x
x
x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x
x x x x x x
x x x
x x
x x x
x x x
3 2 2 2 4 2 2 3 2 4 2 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 4 2 2 3 2 3 7
Journal of Proteome Research • Vol. 5, No. 12, 2006 3283
research articles
Forner et al.
Table 1 (Continued) organelleb
protein name
IDc
G
T
peptidesd
Mitochondria/Cytoplasm/Nucleus Mitochondria/microsomes/peroxisomes Extracellular
Protein DJ-1 Long-chain-fatty-acidsCoA ligase 5 1110039B18Rik protein 17-beta hydroxysteroid dehydrogenase, 1 Adipocyte-derived leucine aminopeptidase Agmat protein Dehydrogenase/reductase SDR family, 8 Fibrinogen, gamma chain Hemoglobin, beta adult major chain Hypothetical protein Lysophosphatidic acid phosphatase 6, 1 Pancreatic alpha-amylase precursor Pregnancy zone protein Serotransferrin Serum paraoxonase/arylesterase 1 Protein CXorf33 homologue Alpha-N-acetylgalactosaminidase Cathepsin B precursor Cathepsin H precursor Dipeptidyl-peptidase II precursor LAMP-2A Lysosomal acid lipase/cholesteryl esterase Lysosomal alpha-glucosidase precursor Lysosomal protective protein precursor Mannosidase, beta A Tripeptidyl-peptidase I
Q99LX0 Q8JZR0 Q8R237 Q8VCR2-1 Q9EQH2 Q8R0Z1 Q9EQ06-1 Q8VCM7 Q9CY10 Q8C717 Q8BP40-1 P00688 Q6PEM2 Q921I1 P52430 Q78IK4 Q9QWR8 P10605 P49935 Q9ET22 P17047-1 Q9Z0M5 P70699 P16675 Q8K2I4 O89023
x x x x x x x x x x x x x x x x x x x x x x x x x x
x
2 10 2 4 2 2 2 3 2 5 4 4 2 4 4 3 4 3 3 2 2 5 3 3 2 3
Extracellular/Mitochondria Lysosomes
x
x x
x
a These proteins were identified by mass spectrometry uniquely in the crude mitochondrial preparation versus the G- or the T-treated mitochondria (marked with an x in the G column if not identified in the G sample, marked with an x in the T column if not identified in the T sample). Protein sequences from the crude mitochondrial sample were blasted against a database composed of either the G- or the T-treated samples protein sequences. Sequence alignments below 80% were considered indicative of the protein presence uniquely in the crude mitochondrial sample. b Subcellular localization according to the SwissProt and GO annotations. c Swiss-Prot or NCBI IDentifier. d Number of unique peptides (see Experimental Section for explanation).
were removed by the proteolysis. This is in agreement with the reported presence of translationally active ribosomes on the surface of mitochondria.33-35 Out of several hundred identified mitochondrial proteins, 25 appear in the crude but not the protease-purified fraction (Table 1). These removed proteins might either be associated with the mitochondrial outer membrane or be membrane proteins partially exposed to the cytosolic side. In the latter case, the protein would seem to disappear if it is only identified by peptides from its cytosolic domains. Crude Mitochondria Treated with Trypsin Compared with Mitochondria Purified by Density Gradient Centrifugation. In a separate experiment, we investigated the extent to which the protease treatment of crude mitochondria was comparable to the density centrifugation gradient in terms of removing nonmitochondrial proteins. First, we compared the proteins identified in the crude mitochondria with those identified in a density gradient-purified mitochondria by performing a Blast analysis of the protein sequences in the crude sample against a database composed of the sequences identified in the gradient-treated sample (Table 1). Overall, we identified 408 proteins in the trypsin-treated and 272 in the density gradienttreated mitochondria with very high confidence (see Experimental Section). The subcellular localization of the proteins that were uniquely identified in the protease-treated sample compared to the density centrifugation gradient sample is shown in Figure 4. A majority of these proteins are known residents of the mitochondria, followed by cytoplasm and endoplasmic reticulum proteins. This result implies that the density centrifugation gradient, though more effective in removing contaminant proteins, can also lose genuine mitochondrial proteins. We compared the number of sequenced peptides for the proteins that were identified in both samples and found them to be similar (Supporting Information Figure 3284
Journal of Proteome Research • Vol. 5, No. 12, 2006
Figure 4. Annotated cellular distribution of proteins identified in the protease-treated fraction but not in the density-gradient centrifugation-purified fraction. The proteins were identified uniquely in the protease-treated mitochondria from liver tissue based on a sequence alignment score below 80% (see text for details). A majority of proteins are annotated as mitochondrial and cytoplasmic. The class “Other membranes” includes microsomes, peroxisomes, and other not further specified subcellular membranes.
3). This indicates that the amount of the mitochondrial proteins that were retained in both purification protocols was roughly the same. Most of the proteins that were removed by the protease treatment were also removed by the density centrifugation gradient. This suggests that the protease treatment acted similarly to the gradient purification, yet through a milder effect (Figure 5a). From the figure, it is quite evident that several mitochondrial proteins were lost upon density-gradient centrifugation, most probably due to partial damage of the mitochondrial structure. This is a clear disadvantage of the
Small-Scale Purification of Mitochondria
Figure 5. Subcellular distribution of proteins depleted by protease treatment (gray line) and by density-gradient centrifugation (black line) compared to the control sample of crude mitochondria. (Upper panel) Reported are the absolute numbers of removed proteins grouped by cellular compartment. (Lower panel) Protein numbers from upper panel were normalized to the number of removed mitochondrial proteins (25 and 61 in the trypsin-treated and in the gradient-treated samples, respectively). M, mitochondria; Cyt, cytoplasm; ER, endoplasmic reticulum; Mem, membranes; non-ann, without specified localization in Swiss-Prot; Nuc, nucleus; Perox, peroxisomes; Secr, secreted.
gradient centrifugation method, which removes more of both the contaminating and the genuine mitochondrial proteins. Indeed, if we normalize the number of proteins from each cell compartment to the number of removed mitochondrial proteins, we find that the two treatments would result in comparable overall performances (Figure 5b). Surprisingly, we observed that some subunits of NADH ubiquinone oxidoreductase (Complex I in the respiratory chain) seemed to be affected by the protease treatment and/or by the density-gradient centrifugation. In the case of protease treatment, the two affected subunits (ESSS, SGDH) are all located in the Iβ sub complex,36 which is the nonmatrix-exposed hydrophobic arm of the complex. Although we cannot exclude that some damage of the mitochondrial membranes occurred, these results would indicate some degree of accessibility of parts of the complex to the protease. Although this observation would require more in-depth studies, we were intrigued by the fact that this huge 45 subunits protein complex had some degree of sensitivity to an external protease. The known function of Complex I is the oxidoreductive transfer of two electrons from NADH to ubiquinone, and this apparently simple reaction seemingly requires 45 known subunits. Recently, new hypotheses have been raised as to the involvement of some of the subunits in other biochemical processes.36,37
research articles Subunits of the subcomplexes IR and Iλ, which are located in the mitochondrial matrix, seemed to be affected by the gradient purification. This is most likely caused by the fact that mitochondria may undergo membrane rearrangements through the density gradient at high centrifugation speeds and/or upon prolonged centrifugation. Overall, the trypsin method presented here is a possible alternative to density-gradient purifications in those situations where the gradients present technical limitations, in particular when a limited amount of biological material is available. Indeed, we found it very difficult to obtain adequate amounts of pure mitochondria after sedimentation on density gradients when starting with
